linux-kernel - Re: [PATCH net v2 2/2] virtio-net: xsk: rx: move the xdp->data adjustment to buf_to

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <ffb4b02e-fadb-4697-b001-7d879ca3b6de@gmail.com>
Date: Tue, 24 Jun 2025 21:58:47 +0700
From: Bui Quang Minh <minhquangbui99@...il.com>
To: Paolo Abeni <pabeni@...hat.com>, netdev@...r.kernel.org
Cc: "Michael S. Tsirkin" <mst@...hat.com>, Jason Wang <jasowang@...hat.com>,
 Xuan Zhuo <xuanzhuo@...ux.alibaba.com>, Eugenio Pérez
 <eperezma@...hat.com>, Andrew Lunn <andrew+netdev@...n.ch>,
 "David S. Miller" <davem@...emloft.net>, Eric Dumazet <edumazet@...gle.com>,
 Jakub Kicinski <kuba@...nel.org>, Alexei Starovoitov <ast@...nel.org>,
 Daniel Borkmann <daniel@...earbox.net>,
 Jesper Dangaard Brouer <hawk@...nel.org>,
 John Fastabend <john.fastabend@...il.com>, virtualization@...ts.linux.dev,
 linux-kernel@...r.kernel.org, bpf@...r.kernel.org
Subject: Re: [PATCH net v2 2/2] virtio-net: xsk: rx: move the xdp->data
 adjustment to buf_to_xdp()

On 6/24/25 17:17, Paolo Abeni wrote:
> On 6/21/25 4:49 PM, Bui Quang Minh wrote:
>> This commit does not do any functional changes. It moves xdp->data
>> adjustment for buffer other than first buffer to buf_to_xdp() helper so
>> that the xdp_buff adjustment does not scatter over different functions.
>>
>> Signed-off-by: Bui Quang Minh <minhquangbui99@...il.com>
>> ---
>>   drivers/net/virtio_net.c | 16 ++++++++++++++--
>>   1 file changed, 14 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
>> index 1eb237cd5d0b..4e942ea1bfa3 100644
>> --- a/drivers/net/virtio_net.c
>> +++ b/drivers/net/virtio_net.c
>> @@ -1159,7 +1159,19 @@ static struct xdp_buff *buf_to_xdp(struct virtnet_info *vi,
>>   		return NULL;
>>   	}
>>   
>> -	xsk_buff_set_size(xdp, len);
>> +	if (first_buf) {
>> +		xsk_buff_set_size(xdp, len);
>> +	} else {
>> +		/* This is the same as xsk_buff_set_size but with the adjusted
>> +		 * xdp->data.
>> +		 */
>> +		xdp->data = xdp->data_hard_start + XDP_PACKET_HEADROOM;
>> +		xdp->data -= vi->hdr_len;
>> +		xdp->data_meta = xdp->data;
>> +		xdp->data_end = xdp->data + len;
>> +		xdp->flags = 0;
>> +	}
>> +
>>   	xsk_buff_dma_sync_for_cpu(xdp);
>>   
>>   	return xdp;
>> @@ -1284,7 +1296,7 @@ static int xsk_append_merge_buffer(struct virtnet_info *vi,
>>   			goto err;
>>   		}
>>   
>> -		memcpy(buf, xdp->data - vi->hdr_len, len);
>> +		memcpy(buf, xdp->data, len);
>>   
>>   		xsk_buff_free(xdp);
>>   
> I'm unsure if this change is in the right direction because it almost
> open-code the existing xsk_buff_set_size() helper - any changes there
> should be reflected here, too.
>
> Also AFAICS xdp->data will now carry a different value, and I guess such
> change is user-visible from the attached xdp program. The commit message
> should at least mentions such fact.

The else path in buf_to_xdp is only triggered when it's called in 
xsk_append_merge_buffer. That's also the reason of the memcpy line 
change in xsk_append_merge_buffer. xsk_append_merge_buffer will allocate 
frag and copy xdp_buff's data to that frag. Previously, after 
buf_to_xdp, the xdp->data does not point to correct place yet so when 
memcpy, we need to copy from xdp->data - vi->hdr_len. With this change, 
we will adjust xdp->data = xdp->data - vi->hdr_len right in buf_to_xdp 
and also adjust other fields (xdp->data_end) to the correct place too. 
So I don't think there is a functional change in this commit.

Thanks,
Quang Minh.